Comparison of effect of cafetière and filtered coffee on serum concentrations of liver aminotransferases and lipids: six months randomised controlled trial.

Objectives: To examine the association of plasma caffeine concentrations during pregnancy with fetal growth and to compare this with relations with reported caffeine intake. Design: Prospective population based study. Setting: District general hospital, inner London. Subjects: Women booking for delivery between 1982 and 1984. Stored plasma was available for 1500 women who had provided a blood sample on at least one occasion and for 640 women who had provided a sample on all three occasions (at booking, 28 weeks, and 36 weeks). Main outcome measure: Birth weight adjusted for gestational age, maternal height, parity, and sex of infant. The exposures of interest were reported caffeine consumption and blood caffeine concentration. Cigarette smoking was assessed by blood cotinine concentration. Results: Caffeine intake showed no changes during pregnancy, but blood caffeine concentrations rose by 75%. Although caffeine intake increased steadily with increasing cotinine concentration above 15 ng/ml, blood caffeine concentrations fell. Caffeine consumption was inversely related to adjusted birth weight, the estimated effect being a 1.3% fall in birth weight for a 1000 mg per week increase in intake (95% confidence interval 0.5% to 2.1%). The apparent caffeine effect was confined to cigarette smokers, among whom the estimated effect was -1.6%/1000 mg a week (-2.9% to -0.2%) after adjustment for cotinine and -1.3% (-2.7% to 0.1%) after further adjustment for social class and alcohol intake. Adjusted birth weight was unrelated to blood caffeine concentrations overall (P = 0.09, but a positive coefficient), after adjustment for cotinine (P = 0.73), or among current smokers (P = 0.45). Conclusions: Smokers consume more caffeine than non-smokers. Blood caffeine concentrations during pregnancy are not related to fetal growth, but caffeine intake is negatively associated with birth weight, with this effect being apparent only in smokers. The effect remains of borderline significance after adjustment for other factors. Prudent advice for pregnant women would be to reduce caffeine intake in conjunction with stopping smoking